What was the first website to hide troll's activity to everyone but the troll himself?
March 25, 2009 10:31 PM
In this article, Clive Thompson says Disqus invented "selective invisibility" -- the moderation technique of hiding the activity of troublemakers and trolls from everyone but the troll themselves. I know this can't be true, mostly because I implemented it on Upcoming.org back in 2003. But I remember the idea floating around long before that, and some old online communities definitely implemented it, but I can't remember which ones. Any old-schoolers remember?
Kuroshin, Slashdot, Plastic, Fark? Maybe even Metafilter? I don't know.
Kuroshin, Slashdot, Plastic, Fark? Maybe even Metafilter? I don't know.
I spoke about it on a panel at SXSW in 2002. I first heard of the concept from Philip Greenspun and his book on building community sites. It's somewhere mentioned in the book, and if I recall I think he called it a bozo filter. If it's not in that exact book (written around 1998) it may have been in something he wrote later, but it'd still be before 2000.
posted by mathowie at 10:41 PM on March 25, 2009
posted by mathowie at 10:41 PM on March 25, 2009
the moderation technique of hiding the activity of troublemakers and trolls from everyone but the troll themselves.
This sounds like it's just a top-down killfile. The roots of this probably extend back to the pre-web era.
posted by pompomtom at 10:43 PM on March 25, 2009
This sounds like it's just a top-down killfile. The roots of this probably extend back to the pre-web era.
posted by pompomtom at 10:43 PM on March 25, 2009
Matt: Thank you! That's the one I was thinking of.
Argyle: Everyone can see comments from anonymous users on Slashdot. It'd only be "selective invisibility" if comments posted by Anonymous Cowards were hidden from the entire community, except for other Anonymous Cowards.
pompomtom: The main difference with a killfile is that it's strictly personal. The point of selective invisibility, or Greenspun's reverse bozo filter, is to discourage trolls from continuing posting by removing their ability to attract attention.
posted by waxpancake at 10:46 PM on March 25, 2009
Argyle: Everyone can see comments from anonymous users on Slashdot. It'd only be "selective invisibility" if comments posted by Anonymous Cowards were hidden from the entire community, except for other Anonymous Cowards.
pompomtom: The main difference with a killfile is that it's strictly personal. The point of selective invisibility, or Greenspun's reverse bozo filter, is to discourage trolls from continuing posting by removing their ability to attract attention.
posted by waxpancake at 10:46 PM on March 25, 2009
Something Awful had this at least in 2002 but possibly earlier. It's called a "hellban" there and was employed secretly by the administrators.
posted by shii at 11:13 PM on March 25, 2009
posted by shii at 11:13 PM on March 25, 2009
I first heard about it in mathowie's SXSW panel, 2002, but I had heard of similar things before.
I have a very vague memory of something like this, maybe, existing in one of the variants of the Citadel BBS software back in the Dark Ages. Although I was the maintainer of one variant and the operator of a BBS running it for a while, I don't remember much from back then.
The thing about community management strategies like this is that the more you talk about them, the less useful they become, so admins are secretive about them. I'm sure every major successful community has features they don't talk about.
My favorite similar tactic: Selective downtime, where the troll finds that the website is down (or really slow) quite often. Not all of the time, because that would tip them off. Trolls are impatient by nature, so they eventually find a more reliable forum to troll.
posted by mmoncur at 11:28 PM on March 25, 2009
I have a very vague memory of something like this, maybe, existing in one of the variants of the Citadel BBS software back in the Dark Ages. Although I was the maintainer of one variant and the operator of a BBS running it for a while, I don't remember much from back then.
The thing about community management strategies like this is that the more you talk about them, the less useful they become, so admins are secretive about them. I'm sure every major successful community has features they don't talk about.
My favorite similar tactic: Selective downtime, where the troll finds that the website is down (or really slow) quite often. Not all of the time, because that would tip them off. Trolls are impatient by nature, so they eventually find a more reliable forum to troll.
posted by mmoncur at 11:28 PM on March 25, 2009
I can't find the reference in that particular Greenspun book, though he does talk about a similar idea, how he modified ACS to return errors for problematic users.
posted by waxpancake at 11:29 PM on March 25, 2009
"The result? Martin got frustrated and went away. Since I'd never served him a 'you've been shut out of this community' message, he didn't get angry with me... Don't tell users that you hate them. Just program your server so that it can pretend to be broken."Similar idea, but not the same as pretending that the troll's comments are being posted and displayed to the rest of the community. In my experience, that's even more effective than returning an error.
posted by waxpancake at 11:29 PM on March 25, 2009
Here's someone else who remembers this on citadel BBSs - the feature was called, for reasons I can't comprehend, "Coventry". Also "Ghost posting". He also refers to an even more devious version, where any useful messages posted by the "ghost" can be promoted to visible status, making it even less obvious to them that they're invisible.
I'm 90% certain now that I had this feature in Fnordadel for the Atari ST in about 1987, but I don't have any way to access the source code without hauling the ST out of storage.
Also, Greenspun seems to have had my "selective downtime" idea before I did, so I won't take the credit...
posted by mmoncur at 11:45 PM on March 25, 2009
I'm 90% certain now that I had this feature in Fnordadel for the Atari ST in about 1987, but I don't have any way to access the source code without hauling the ST out of storage.
Also, Greenspun seems to have had my "selective downtime" idea before I did, so I won't take the credit...
posted by mmoncur at 11:45 PM on March 25, 2009
It was common practice on some usenet groups as early as 1991 for people to publish their killfiles. These killfiles often suppressed the display of the common group trolls for whomever used them in their newsreader.
The trolls could still post whatever they wanted, but anyone in the know wouldn't have to see their dreck.
See for example jwz's post to rec.music.misc on Mar 21, 1991:
For entertainment purposes only; PLEASE, no wagering.
Here is my GNUS kill file for rec.music.{misc|cd}.
Use it in good health.
posted by zippy at 11:45 PM on March 25, 2009
The trolls could still post whatever they wanted, but anyone in the know wouldn't have to see their dreck.
See for example jwz's post to rec.music.misc on Mar 21, 1991:
For entertainment purposes only; PLEASE, no wagering.
Here is my GNUS kill file for rec.music.{misc|cd}.
Use it in good health.
posted by zippy at 11:45 PM on March 25, 2009
the feature was called, for reasons I can't comprehend, "Coventry"
To be 'sent to Coventry' is to be ostracised for some transgression.
It's the perfect name for the feature.
posted by pompomtom at 11:49 PM on March 25, 2009
To be 'sent to Coventry' is to be ostracised for some transgression.
It's the perfect name for the feature.
posted by pompomtom at 11:49 PM on March 25, 2009
Ah, thanks for explaining that. Now I have another bit of British slang with which to annoy my fellow Americans.
posted by mmoncur at 11:52 PM on March 25, 2009
posted by mmoncur at 11:52 PM on March 25, 2009
On the Citadel BBSes I frequented in the late '80s, it was referred to as the "twit bit", and the disliked user would sometimes be said to have been "twitted".
posted by hattifattener at 11:56 PM on March 25, 2009
posted by hattifattener at 11:56 PM on March 25, 2009
I remember the "twit bit", but as I recall it wasn't quite as subtle as "sending them to Coventry" - their messages went to the Trash room rather than being posted (but invisible) in the room they tried to post in.
posted by mmoncur at 12:03 AM on March 26, 2009
posted by mmoncur at 12:03 AM on March 26, 2009
It's possible that I'm misremembering terms… but I do remember the described behavior (user's posts become invisible to everyone except themselves and sysops, so that the user doesn't know they've been silenced) and I think it was referred to as the twit bit (a bit set on the user's account).
I suppose it's also possible that the terminology was different in different lineages of the BBS software; there were quite a few.
posted by hattifattener at 12:12 AM on March 26, 2009
I suppose it's also possible that the terminology was different in different lineages of the BBS software; there were quite a few.
posted by hattifattener at 12:12 AM on March 26, 2009
I pulled down a handful of early Citadels' source code to look. The earliest Citadel (v 2.10, circa 1981) doesn't seem to have the feature. Citadel-86, from a few years later, has a "twit bit", but it's more as mmoncur describes. GremCit (c.1987), one of the earliest I remember using — as the archive page notes, "This code served as the root for almost all subsequent Citadel development in the Seattle area" — has a "problem user" bit, which does sound like what waxpancake is talking about: posts by such users are only visible to "aides" (aides are non-sysops who've been given some extra admin-type privs), but can also have their general visibility toggled on and off on a per-post basis by aides. It looks like that feature migt have come from Stonehenge (c. '85-'86), but its archive uses some compression methods that modern unzip implementations don't seem to support.
posted by hattifattener at 12:42 AM on March 26, 2009
posted by hattifattener at 12:42 AM on March 26, 2009
The roots of this probably extend back to the pre-web era.
Definitely. I was doing this on a local BBS by 1991-1992: I realized that you could limit the users on forums you had set up (the ADMIN ones, for example, just had sysadmins). It was easy to then to create a forum on a topic that would enrage a known troll and just set it so the only users who could see it were the troll any myself. I would post a single item and then watch him try to rabble-rouse and wonder aloud why no one was taking the bait. He would then allude to it in other discussions, and other users would scratch their heads and wonder what he was on about, so it was not a well-known tactic then.
posted by ricochet biscuit at 6:12 AM on March 26, 2009
Definitely. I was doing this on a local BBS by 1991-1992: I realized that you could limit the users on forums you had set up (the ADMIN ones, for example, just had sysadmins). It was easy to then to create a forum on a topic that would enrage a known troll and just set it so the only users who could see it were the troll any myself. I would post a single item and then watch him try to rabble-rouse and wonder aloud why no one was taking the bait. He would then allude to it in other discussions, and other users would scratch their heads and wonder what he was on about, so it was not a well-known tactic then.
posted by ricochet biscuit at 6:12 AM on March 26, 2009
" It looks like that feature might have come from Stonehenge (c. '85-'86), but its archive uses some compression methods that modern unzip implementations don't seem to support."
I didn't have any trouble opening said Stonehenge archive (it's a ZIP 1.0 header and files archived with Shrink and Implode methods and InfoZip handles it just fine) but (a) there's no source in it and (b) the feature in question is not documented in either of sysop.doc or aide.doc.
I concur that it's almost certainly going to be a Citadel-derived BBS that represents the earliest actual implementation of the feature as described. Nailing down which one, though, is a nightmare: Citadel has more forks than an army mess hall.
posted by majick at 6:48 AM on March 26, 2009
I didn't have any trouble opening said Stonehenge archive (it's a ZIP 1.0 header and files archived with Shrink and Implode methods and InfoZip handles it just fine) but (a) there's no source in it and (b) the feature in question is not documented in either of sysop.doc or aide.doc.
I concur that it's almost certainly going to be a Citadel-derived BBS that represents the earliest actual implementation of the feature as described. Nailing down which one, though, is a nightmare: Citadel has more forks than an army mess hall.
posted by majick at 6:48 AM on March 26, 2009
The Telnet BBSes (running Citadel) that I adminned in the late 80s and early 90s had a feature called "Twit".
The "twitted" user could still post, but only they could see their posts. You also had the option of having all of their posts automatically moved to a separate forum.
Old users of Quartz BBS: Represent!
posted by DWRoelands at 7:01 AM on March 26, 2009
The "twitted" user could still post, but only they could see their posts. You also had the option of having all of their posts automatically moved to a separate forum.
Old users of Quartz BBS: Represent!
posted by DWRoelands at 7:01 AM on March 26, 2009
You guys are awesome.
posted by waxpancake at 10:08 AM on March 26, 2009
posted by waxpancake at 10:08 AM on March 26, 2009
majick: Ah, OK, it's apparently just that my copy of unzip was compiled without patent-encumbered methods. Anyway, "problem users" are mentioned in several of the files I can extract (AIDE.DOC, AIDE.HLP, and SYSOP.DOC), with enough description that I think it's the feature waxpancake is looking for (the one I and DWRoelands remember as the twit bit, despite cit86's different use of the term).
Also: family tree. Bob Perigo's Babel was one of the first BBSes I ever called, I think.
posted by hattifattener at 10:36 AM on March 26, 2009
Also: family tree. Bob Perigo's Babel was one of the first BBSes I ever called, I think.
posted by hattifattener at 10:36 AM on March 26, 2009
Any user interface related to messaging in the small-community space (less than 200+ people interacting) was generally invented in 1960-1985, and anything in the larger community, likely in 1985-1992. Anyone indicating anything after that is delusional.
posted by jscott at 12:15 PM on March 26, 2009
posted by jscott at 12:15 PM on March 26, 2009
Slashdot's Anonymous Cowards get their comments seen; they're just a click or two harder to get to by default. We don't hide the fact that A.C. comments start at a lower score, there's no trickery going on there.
This is kind of the opposite of what was requested, but I recall there was some jerk back in the day (late 80s?) who lurked in a MU* room and waited until just a few seconds before New Year's to mute each player to everyone except themselves. He then posted a log of 30+ people who were all chatting and having a good time right up until the big moment at which point they started wishing each other happy new year and not seeing any postings by anyone else. I think the intention was to show what losers MU* players were (because it took them several minutes to realize what was going on), but to my mind the hacker was just being petty. Can't remember any terms to search on to find this, sorry.
posted by Jamie McCarthy at 5:46 PM on March 26, 2009
This is kind of the opposite of what was requested, but I recall there was some jerk back in the day (late 80s?) who lurked in a MU* room and waited until just a few seconds before New Year's to mute each player to everyone except themselves. He then posted a log of 30+ people who were all chatting and having a good time right up until the big moment at which point they started wishing each other happy new year and not seeing any postings by anyone else. I think the intention was to show what losers MU* players were (because it took them several minutes to realize what was going on), but to my mind the hacker was just being petty. Can't remember any terms to search on to find this, sorry.
posted by Jamie McCarthy at 5:46 PM on March 26, 2009
some old online communities definitely implemented it, but I can't remember which ones. Any old-schoolers remember?
in 1983, here in Seattle, I was supervising the development of Stonehenge BBS by David Bonn at a startup company - the company failed soon after, but some of the software projects survived with their authors in control.
the twit filter/bit was our solution to flamers - the idea was sparked by the Monty Python upper cass twit skit
not sure whether it pre-existed in some other form
steve estvanik / "yngvi"
posted by cascoly at 12:41 PM on April 18, 2009
in 1983, here in Seattle, I was supervising the development of Stonehenge BBS by David Bonn at a startup company - the company failed soon after, but some of the software projects survived with their authors in control.
the twit filter/bit was our solution to flamers - the idea was sparked by the Monty Python upper cass twit skit
not sure whether it pre-existed in some other form
steve estvanik / "yngvi"
posted by cascoly at 12:41 PM on April 18, 2009
We built this into a marcom campaign site for Levi's back in 1997. It was a "[blank] wore them" billboard, where you could put your name in and add it to a huge page of thousands of people. "Calvin Klein wore them. Carrie Southerland wore them. Chinese Sweatshoppers wore them." etc.
When you submitted a name to the site you'd immediately see it posted on the combined wall, but would only be visible to others after approval. If it was rejected, the posted (ID'd by cookie) would continue to see it on their wall, and they'd have the added joy of thinking they were the only person to be so witty.
posted by kfury at 5:37 PM on April 29, 2009
When you submitted a name to the site you'd immediately see it posted on the combined wall, but would only be visible to others after approval. If it was rejected, the posted (ID'd by cookie) would continue to see it on their wall, and they'd have the added joy of thinking they were the only person to be so witty.
posted by kfury at 5:37 PM on April 29, 2009
« Older IE6 is frakking up the sidebar of my website. | how much will a large dog cost per year Newer »
This thread is closed to new comments.
posted by Argyle at 10:41 PM on March 25, 2009